Re: GEOM_GPT on i386?

From: Brooks Davis <brooks_at_one-eyed-alien.net>
Date: Thu, 9 Sep 2004 15:51:22 -0700
On Fri, Sep 10, 2004 at 12:42:27AM +0200, Harald Schmalzbauer wrote:
> Am Freitag, 10. September 2004 00:37 schrieb Brooks Davis:
> > On Fri, Sep 10, 2004 at 12:31:35AM +0200, Harald Schmalzbauer wrote:
> > > Am Freitag, 10. September 2004 00:28 schrieb Brooks Davis:
> > > > On Fri, Sep 10, 2004 at 12:17:07AM +0200, Harald Schmalzbauer wrote:
> > > > > Dear GEOM coders,
> > > > >
> > > > > I saw that some time ago GPT was enabled in GENERIC.
> > > > > Is it possible to create GPT labels on i386 and if so, how do I do
> > > > > that? The GPT man page tells about ia64 (only?)
> > > >
> > > > The gpt(8) command should work on any architecture it is installed on
> > > > (!sparc64).
> > >
> > > Thank you, but this doesn't neccessarily mean that I can use GPT to boot
> > > my i386 5.3 box to get rid of the 8-label Limit, does it?
> > > Has anybody succesfully changed to GPT on i386 (by `gpt migrate
> > > /dev/ad0`)?
> >
> > As far as I know, there is no boot support for GPT yet.  That's
> > something I would like to see, but it's not there yet.  Which feature(s)
> > of GPT do you need?
> 
> Hmm, can I make use of GPT without having boot support? Perhaps you
> have a link handy which illustrates the GPT on disk and the BIOS
> interactivity. Is it possible to have both, MBR and GPT? I can't
> imagine.  My particular problem is that I want to have every jail in
> it's own label, but then I have to slice my disk (array), because
> I have roundabout 20 jails, which makes it impossible to rearrange
> label-sizes.

While GPT was intended to be the only label on the disk, GEOM does not
have any such restrictions.  You can place any label on any geom.  My
suggestion would be to allocation a slice or bsdlabel partition for the
GPT.  Then use GPT to chop that up to meet the demands of the jails.
For example, you could have:

da0s1a		/
da0s1b		swap
da0s1d		/var
da0s1e		/tmp
da0s1f		/usr
da0s2		<GPT partioned jail space>

da0s2 could also be da0s1g.

-- Brooks

-- 
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4

Received on Thu Sep 09 2004 - 20:49:24 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:11 UTC